# Docs MCP Server

> Connect Claude, Cursor, VS Code and other AI tools to the Stacksync documentation

Source: https://docs.stacksync.com/docs-for-ai/mcp-server

The Stacksync docs MCP server lets AI assistants search and read the documentation, and ask questions about it, while they help you. It uses the [Model Context Protocol](https://modelcontextprotocol.io) over streamable HTTP. It's public and needs no account or API key.

```
https://docs.stacksync.com/api/mcp
```

<Callout type="info">
  This server gives access to the **documentation** only. It can't read or change your Stacksync workspace, syncs or data.
</Callout>

### Connect your AI tool

<Tabs items="[&#x22;Claude Code&#x22;,&#x22;Claude&#x22;,&#x22;Cursor&#x22;,&#x22;VS Code&#x22;,&#x22;Other&#x22;]">
  <Tab value="Claude Code">
    Run:

    ```bash
    claude mcp add --transport http stacksync-docs https://docs.stacksync.com/api/mcp
    ```
  </Tab>

  <Tab value="Claude">
    In Claude (web or desktop), open **Settings → Connectors**, click **Add custom connector**, name it `Stacksync Docs` and paste the server URL:

    ```
    https://docs.stacksync.com/api/mcp
    ```
  </Tab>

  <Tab value="Cursor">
    Add the server to `.cursor/mcp.json` in your project, or to `~/.cursor/mcp.json` for every project:

    ```json
    {
      "mcpServers": {
        "stacksync-docs": {
          "url": "https://docs.stacksync.com/api/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab value="VS Code">
    Add the server to `.vscode/mcp.json` in your project:

    ```json
    {
      "servers": {
        "stacksync-docs": {
          "type": "http",
          "url": "https://docs.stacksync.com/api/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab value="Other">
    Any MCP client that supports streamable HTTP servers can connect with the URL `https://docs.stacksync.com/api/mcp`. No authentication is needed.
  </Tab>
</Tabs>

### Tools

| Tool         | What it does                                                                                                                         | Input                                                                          |
| ------------ | ------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------ |
| `search`     | Searches the docs and returns the best matching pages, with their URLs, Markdown URLs, breadcrumbs and matching sections             | `query`, optional `limit` (1–20, default 8)                                    |
| `get_page`   | Returns the full Markdown of a page                                                                                                  | `url`: a page URL, its `.md` URL or its path, e.g. `/two-way-sync/quick-start` |
| `list_pages` | Returns the index of every page ([`llms.txt`](/docs-for-ai#llmstxt))                                                                 | none                                                                           |
| `ask`        | Answers a question in plain language from the docs, with links to the pages it used. Slower than `search`; best for how-to questions | `question`                                                                     |

A typical agent calls `search`, then `get_page` on the results it needs. `ask` is the same assistant as **Ask AI** on this site, and is also available without MCP through the [Ask API](/docs-for-ai/ask-api).

### Example prompts

Once connected, ask your AI tool things like:

* "Using the Stacksync docs, how do I set up a two-way sync between Salesforce and Postgres?"
* "Search the Stacksync docs for the Salesforce OAuth error and tell me how to fix it."
* "Read the Stacksync API Proxy quick start and write a Python script that lists HubSpot contacts through it."